WO2008029057A1

WO2008029057A1 - Method of integrated routing in a federation of personal networks

Info

Publication number: WO2008029057A1
Application number: PCT/FR2007/051869
Authority: WO
Inventors: Usman Javaid; Tinku Mohamed-Rasheed; Djamal-Eddine Meddour; Toufik Ahmed
Original assignee: France Telecom
Priority date: 2006-09-05
Filing date: 2007-09-04
Publication date: 2008-03-13

Abstract

This method of routing is intended for federating a plurality of networks (60...68;70...76), each network comprising its own topology of nodes of terminals which is represented by an associated network routing table (505) and the plurality of networks comprising at least two gateways interconnecting two networks. The method comprises a step of binding at least one network to the federation by implementing a predetermined consent protocol and a step of formulating a federation routing table (506) comprising a federation topology dependent on the actual topologies of the networks bound to the federation and of choosing common resources operated by said networks. The step of formulating a federation routing table (506) takes into account messages exchanged during the binding step, without resorting to subsequent message exchanges. Figure 5.

Description

Routing method integrated in a federation of personal networks.

The present invention relates to a method of pooling resources of at least one terminal of each of a plurality of networks.

It is placed within the framework of personal networks. Personal networks are a recent extension of geographically limited Personal Area Networks (PANs).

In fact, NAPs are networks of limited size, located geographically. The terminals making up the network are typically spaced a few meters apart. Only one communication technology is used in a PAN, for example Wi-Fi or Bluetooth (registered trademarks).

Personal networks are freed from geographical limitation by proposing to be divided into sectors, each sector being substantially the equivalent of a NAP, except that a sector may comprise several communication technologies.

Today's society tends towards the sharing of information and data. In this context, it is necessary to provide mechanisms for the pooling of resources between personal networks. This pooling is later called "federation of personal networks".

Mechanisms for pooling resources and federation architectures of personal networks are generally known.

However, the routing mechanisms currently available which are those of personal network type (Wl-Fl or Bluetooth) or those of type interconnection network (UMTS for example) are heavy to implement in the context of a federation of personal networks, especially within the framework of a local federation, called spontaneous or direct network between personal networks.

Indeed, it appears that if the number of personal networks to integrate in the federation is important, the signaling volume required for routing with a conventional routing mechanism is prohibitive. The invention therefore aims to provide a routing method for the federation of a plurality of networks, particularly personal, which is fast and uses a limited number of signaling resources.

For this purpose, the subject of the present invention is a routing method intended for the federation of a plurality of networks, the plurality of networks comprising at least two interconnection gateways between two networks, each network comprising a terminal topology represented by an associated network routing table and network operable resources, the method comprising:

a step of joining at least one network to the federation by implementing a predetermined approval protocol,

a step of developing a federation routing table by said at least one network, a federation routing table comprising a federation topology according to the clean topologies of the networks having joined the federation and of the choice of common resources operated by said networks, characterized in that the step of developing a federation routing table takes into account messages exchanged during the membership step, without recourse to subsequent message exchanges.

Thanks to the invention, the signaling resources required for the federation are limited to the federation topology and are formed in one go at the time of the implementation of the approval protocol. The resources operable by the network are resources of a terminal, data or network resources such as supported radio interfaces, access rights, ....

According to one characteristic of the invention, the method comprises the following steps:

- Develop an invitation message for pooling resources including an identifier of an inviting network,

sending the invitation message from the inviting network to each other guest plurality network, following reception of the invitation message by a guest network, elaborating an acceptance message for the pooling of resources comprising at least the address of a terminal of this guest network,

sending the acceptance message from at least one guest network to the inviting network,

selecting at least one terminal of the inviting network and at least one terminal of the guest network contained in the acceptance message or messages, in order to pool all or part of the resources of these terminals,

sending to the networks of the plurality, a pooling message including the list of selected terminals,

- develop, based on data conveyed by at least part of the invitation, acceptance and pooling messages, a data transfer system allowing the exchange of data between the selected terminals, without recourse to exchanges subsequent messages.

Thanks to the invention, the implementation of the data transfer system is done at the same time as the message exchanges. This saves time: once terminals are selected, connectivity between them is established and pooling is ready for use. Moreover, the implementation is completely transparent for the users.

According to one characteristic of the invention:

at least two networks of the plurality are distant from each other, separated by an interconnection network, and the method comprises the following steps:

configuring, in the interconnection network, a first agent associated with the first remote network and a second agent associated with the second remote network, so that each network and its associated agent can transfer data between them,

finding the address in the interconnection network of another agent from the identifier of the network associated with this other agent,

- carry out the exchange of messages between the two remote networks via the agents, developing the data transfer system by making available to each network the addresses of the selected terminals provided by the acceptance messages, in order to be able to use a routing provided by the interconnection network.

Since the agents are external to the networks, the users of these networks do not have to configure them. Thus, the search for a network is done very easily: the inviting user only has to know an identifier of the remote networks he wishes to invite.

According to another characteristic of the invention: data is exchanged between at least two sectors of a network thanks to the associated agent, the sectors being distant from each other, separated by the interconnection network.

This feature of the agents is already known. The realization of the communication agents between sectors by improving them agents according to the invention, facilitates the implementation of the invention since it has a structure already existing, and therefore close to easy operation easily adaptable.

According to another characteristic of the invention: the method comprises the step of transferring to each agent, the list of selected terminals.

This feature adds security, since the agent may refuse to grant access to the network to unlisted terminals.

According to another characteristic of the invention: networks, forming at least one subset of the plurality of networks and comprising the inviting network, are connected directly to each other, and the step of developing a system for transferring data comprises the development, by each network of the set, of a routing table among the networks of the set, from the conveyed data consisting of the routing in the subset of the invitation and acceptance messages.

This feature provides a simple way to transfer data between adjacent (directly connected) networks since the development of routing tables between terminals is well known and it is therefore possible to take advantage of what already exists to adapt it to the case of routing between networks.

According to another characteristic of the invention: the invitation message is sent to the networks of the set step by step starting from the inviting network.

This feature provides a simple and systematic way to send the invitation message.

According to another characteristic of the invention:

the invitation message includes a list of networks of the set,

the inviting network sends the invitation message to the networks of the list directly connected to the inviting network,

- After receiving the invitation message by a network, it is removed from the list, then the network sends the modified invitation message to the networks of the modified list directly connected to it.

The endless loops of message transmission are thus avoided.

According to another characteristic of the invention: the pooling message comprising the list of selected terminals is stored in each terminal of each network of the set, forming a gateway with another network of the set.

The security is thus improved, since the gateways can prevent access to the network to unlisted terminals.

According to another characteristic of the invention:

each terminal of each network of the plurality establishes a local routing table, in which each directly connected terminal is associated with the IP address of this terminal and at least one MAC address of a network interface to which this terminal is connected. directly.

Thus, it is possible to manage networks having heterogeneous connections, that is to say different technologies. Indeed, the MAC address of a network interface indicates the type of technology of the interface. The invention also relates to a software for executing the method as defined above.

The subject of the invention is also a network federation system comprising a plurality of networks, each network comprising a terminal nodes own topology represented by an associated network routing table and the plurality of networks comprising at least two gateways. interconnection between two networks, characterized in that it comprises means for developing a federation routing table, for at least one network having joined the federation according to a predetermined approval protocol, a federation routing table comprising a federated topology according to the proper topologies of the networks having joined the federation and the common resource choices operated by said networks, and in that the means for producing said federation routing table are configured to develop the table routing from messages exchanged during the execution of the approval protocol, without course to subsequent message exchanges.

The invention also relates to a terminal intended to be selected in a federation of networks of a plurality and belonging to a network of the plurality characterized in that it comprises a routing engine comprising a module for discovering adjacent terminals, an adjacent network discovery module, a federation management module, an authentication module, a routing table for the network to which the terminal belongs, and a routing table for the federation of networks to which said network belongs terminal.

The invention will be better understood in the light of the description which follows, given solely by way of example and with reference to the appended drawings, among which:

FIG. 1 represents a personal network,

FIG. 2 represents three configurations of federations of personal networks, FIG. 3 is a functional representation of the communication interface present on each terminal of a personal network,

FIG. 4 represents a sector of a personal network,

FIG. 5 represents a so-called local federation of personal networks, and

FIG. 6 represents a so-called remote network federation.

Referring to Figure 1, a personal network, designated by general reference 10, is shown.

In general, a personal network is linked to a person, that is to say that the personal network includes the terminals belonging to this person, hereinafter called "the owner". The personal network is characterized by an identifier.

Thus, the personal network 10 consists of a plurality of terminals of an owner, distributed in three sectors 14, 16, 18.

A sector is a group of terminals close to each other and in direct communication. Generally, the distance between the terminals of a sector is of the order of one meter or ten meters.

The expression "in direct communication" means that two terminals are connected using a single technology and that they each have respective interfaces to communicate using this technology.

In the context of the invention, the technologies for short-distance communication between terminals of the same sector are wireless technologies, such as for example: Wi-Fi, Bluetooth ...

The different sectors 14, 16, 18 of the personal network 10 are interconnected via one or more interconnection networks 24. In the example illustrated, only one interconnection network 24 is present. This is the Internet network.

Sector 14 includes the terminals that the owner has on him, namely a 14-ι mobile phone and a 14 ₂ laptop. These ter- males 14 _1; 14 ₂ each include a respective Wi-Fi interface through which they communicate directly with each other.

Sector 16 corresponds to the "domestic" terminals, that is to say located in the home of the owner.This sector 16 includes a personal computer 16-, connected to the Internet network 24 via an ADSL modem, a printer 16 ₂ connected to the computer via a USB link, and a camera ₃ connected to the computer via a Wi-Fi connection.

Sector 18 is located at the owner's workplace. This sector 18 includes an 18-ι work computer connected to the Internet network 24 by a Wi-Fi wireless network card and having a Bluetooth interface, an 18 ₂ cellular telephone having a Bluetooth interface, and a personal digital agenda or PDA 18 ₃ also equipped with a Bluetooth interface.

Thus, a terminal may have multiple connections using different technologies. For example, the work computer 18i has a Wi-Fi connection and a Bluetooth connection.

Each of the sectors 14, 16, 18 therefore comprises at least one terminal having a connection with the interconnection network 24. This terminal forms a gateway with the interconnection network 24. This type of terminal will be called thereafter gateway interconnection. For example, the mobile phone 141 is able to use UMTS technology to connect to the Internet network 24.

In order to allow communication between the different sectors 14, 16, 18 of the personal network 10, a database 26, also called personal network agent, is provided in the interconnection network 24. This agent 26 stores the addresses of the different interconnection gateways of all sectors 14, 16, 18 of the network 10, in order to properly route the messages to each sector.

In the illustrated example, the agent 26 stores the address of the mobile phone 14-1, the personal computer 16i and the work computer 18-ι.

In addition, each of the interconnection gateways, here 14-, 16-, and 18i, knows the address of the personal network agent 26, in order to transmit to this agent 26 the messages addressed to a user. sector other than his own. In the remainder of the description, for the sake of clarity, the personal networks will comprise only one sector.

With reference to FIG. 2, three federations of personal networks, designated respectively by the references 30-ι, 30 ₂ and 30 ₃ , are represented.

A federation is a pooling of terminals of at least two personal networks, and more generally the resources provided by these terminals. It should be noted that all the terminals of a personal network are not necessarily provided to the federation.

Six personal networks are concerned by the 30-ι, 3O ₂ and 3O ₃ federations. They are respectively designated 32 _1; 32 ₂ , 32 ₃ , 32 ₄ , 32 ₅ and 32 ₆ .

Two personal networks 32 _1; 32 ₂ form the first federation 30-ι. They are interconnected by an interconnection network 34, here again the Internet network. This type of federation will be called later remote federation.

In this type of federation, the communications between the two personal networks 32 _1; 32 ₂ go through the Internet 34. Each of the personal networks 32 _1; 32 ₂ therefore comprises an interconnection gateway, which has a connection with the Internet 34.

Three other personal networks 32 ₃ , 32 ₅ , 32 ₆ form the second 3O ₂ federation. Personal networks 32 ₃ , 32 ₅ , 32 ₆ are connected directly to each other. This means that each of them includes at least one terminal forming a gateway with one of the other personal networks. This terminal, hereinafter called local gateway, has a direct connection with a terminal of one of the other personal networks. This type of federation will be called later local federation.

The third 3O ₃ federation brings together the three personal networks 32 ₃ , 32 ₄ , 32 ₆ . This is a mixture of the two previous configurations: the network 32 ₃ is connected directly to the network 32 ₆ , but it is connected to the network 32 ₄ via the Internet 34. This type of federation will be called by the Hybrid Federation. Referring to Figure 3, we will now describe, from a functional point of view, the communication interface present in each terminal of a personal network. For this purpose, a terminal of a personal network is considered.

Its interface has a layered architecture. FIG. 3 thus represents a data link layer 40 (called "layer 2"), a network layer 42 (called "layer 3") and application layers 44 (layers greater than 3).

The data link layer 40 manages communications between the considered terminal and an adjacent terminal, that is to say directly connected by a physical medium. It includes a module 46 for monitoring the Quality of Service of each network interface 48-ι, 48 ₂ ... present on the terminal in question.

Network layer 42 manages end-to-end communications between terminals. It comprises a routing engine 50, providing routing tables as will be explained later, and a routing module for selecting an optimal route for the transfer of data using the routing tables determined by the motor 50 and Quality of Service information provided by the link layer 40, and by the upper layers 44.

The routing engine 50 comprises a module 5O ₁ for discovering adjacent terminals, a module 50 ₂ for discovering adjacent personal networks. It further comprises a module 5O ₃ federation management and a module 5O ₄ authentication. The routing engine 50 finally comprises a routing table 5O ₅ for the personal network to which the terminal belongs, as well as a routing table 5O ₆ for the federation of personal networks which includes the terminal in question.

The upper layers 44 comprise an application level of service quality module 52, a service detection module 54, a certification management module 56, and a profile management module 58 able to implement an approval protocol. of federation. We will now describe the operation of the routing engine 50, to determine the topology of the sector to which the terminal considered.

The operation begins with a step 100 of discovery, by the module of discovery of adjacent terminals 5O ₁ , terminals adjacent to the terminal in question. This step is for example implemented using the reception and transmission of beacon signals specific to the access technology at the level of the MAC layer. As a variant, the step is performed at the level of the link layer 42 by sending a HELLO message through each of the network interfaces 48-ι, 48 ₂ , ...

In the remainder of the description, it will be assumed that HELLO messages are used.

There follows a step 102 of receiving responses sent by the adjacent terminals.

The authentication module 50 ₄ then implements an authentication step 104 with the adjacent terminals that have responded. This step typically involves different message exchanges between the terminal in question and the adjacent terminals. For each adjacent terminal, it is thus possible to determine if this terminal belongs to the same personal network, if it belongs to another personal network, or if it does not belong to any personal network, but to another type of network such as than the Internet.

In the last two cases, the terminal in question is therefore a gateway, respectively a local gateway or an interconnection gateway.

Then, during a step 106, the preceding information is grouped together in a local routing table, that is to say concerning only the adjacent terminals. This local routing table comprises the list of network interfaces of the terminal in question and, associated with each network interface, its physical address MAC and the IP address of the adjacent terminal connected to this interface. This table contains a field indicating whether the adjacent terminal 100 belongs to the same personal network, to another personal network, or to another type of network.

During a step 108, the local routing table is sent to all the adjacent terminals of the personal network, but not outside the personal network. This step is reproduced periodically.

As the adjacent terminals also perform this step 108, the terminal considered receives a set of routing tables, from which the routing engine 50 updates the routing table 5O ₅ of the personal network.

The routing table 5O ₅ of the personal network thus comprises a list of possible source and destination terminal pairs in the personal network identified by their IP address. For each pair, the routing table 5O ₅ of the personal network includes the IP address of each next adjacent terminal, on a terminal route connecting the source to the destination. The MAC physical address of each network interface to reach this next terminal on the road is also recorded in the table 5O ₅ , as well as the associated cost. The cost is for example the number of terminals to cross using this route.

By way of example, FIG. 4 shows four terminals A, B, C, D forming a sector of a personal network. The terminal B is connected to each of the other three terminals A, C, D by a respective network interface α, γ, δ. The network connections of the terminal B are, in this example, of different types. Moreover, the terminals D and C are also connected to each other. Thus, the local routing table at terminal B is expressed according to the invention:

IPA MACα L

IPC MACγ L

IPD MACD L

The first column is the IP address of the adjacent terminal, the second column is the MAC address of the network interface to which the terminal is connected. adjacent and the third column indicates the type of adjacent terminal, i.e. here local (L).

After exchange of the local routing tables, the routing table 5O ₅ to the terminal B is expressed according to the invention:

IPA IPC IPC MACγ 1

IPA IPC IPD MACδ 2

IPA IPD IPD MACD 1

IPA IPD MACG IPC 2

IPA IPA IPA MACα 1

The first column is the IP address of the source terminal, the second column is the IP address of the destination terminal, the third column is the IP address of the next adjacent terminal on a route to the destination terminal, the fourth column is MAC address of the network interface to which is connected the next adjacent terminal and the fifth column is the associated cost, here the number of terminals to cross to reach the destination.

Due to the fact that the routing table includes both the IP address of the adjacent terminal and the MAC address of the network interface to which it is connected, heterogeneous network interfaces can be used on the same terminal, in a transparent manner. .

Indeed, the MAC address gives the type of interface and it is therefore possible to adapt the data to the type of interface.

It will be noted that the routing table 5O ₅ also makes it possible to route the data to the adjacent networks, since they were discovered in step 104.

With reference to FIG. 5, the formation of a local federation of personal networks will now be described.

In this example, five personal networks 60, 62, 64, 66 and 68 are involved. Local gateways between personal networks are shown in hatched lines. They are referenced 60-ι, 62 _1; 64 _1; 64 _2, 661, 66 _2, 681. The owner of the personal network 60 decides to form a federation of personal networks, for example to launch an online video game, or to share vacation photos. The owner subsequently uses any of the terminals of his personal network 60.

The owner creates, during a step 200, a federation profile.

The federation profile includes an identifier of its creator, that is to say the identifier of the personal network of the creator, a textual description indicating the purpose of the federation, the access control policies, the resources required for the federation, as well as a list of participants identified by the identifier of their personal network. For example, the list of participants consists of personal networks 64, 66, 68.

The owner of the personal network 60 then creates, during a step 202, a participation profile in which he defines in particular the terminals of his personal network 60 all or part of the resources are made available to the federation.

During a step 204, the two profiles, federation and participation, are made available to the local gateways of the creator's personal network, in the example, it is the local gateway 601.

This provision is made for example by copying the profiles in the local gateways.

Alternatively, it is possible to simply define a pointer to each profile stored in another terminal of the personal network.

The federation profile is then distributed to the participants' personal networks. This process is a (special) development of the accreditation protocol.

For this purpose, during a step 206, an invitation message to join the federation, including the federation profile and therefore the list of participants, is transmitted to the personal networks directly connected to the personal network 60, which are present in the list of participants. The routing table 5O ₅ of the personal network 60 is used for this purpose. Similarly, all subsequent data transmissions or messages are routed in a personal network thanks to its routing table 5O ₅ . Thus, in the example of FIG. 5, the request is transmitted to the personal network 64 via the local gateway 6O1.

The following steps are performed by each personal network receiving the invitation message.

The personal network receiving the invitation message presents it to its owner (step 210) using a terminal-owner tracking device. If the owner accepts the invitation, the request is modified by removing from the participant list the identifier of the personal network considered (step 212), and updating (step 214) the federation routing table, to add the personal network from which the request was received, and the gateway to use to reach the request.

In addition, the owner of the personal network creates (step 216) a participation profile indicating the terminals of his personal network all or part of the resources are made available to the federation and transmits (step 218) an acceptance message comprising the participation profile for the personal network 60 of the creator.

During a step 220, the modified invitation message is transmitted to personal networks directly connected and still present in the list of participants.

In addition, the personal network considered receives the acceptance messages from the personal networks to which it transmitted, in step 218, the modified invitation message. It then uses, during a step 222, these acceptance messages to update its federation routing table. The federation routing table is updated by the gateway.

Thus, at the end of the exchange of profiles between the different personal networks of the participants and the personal network of the creator, each personal network has a complete federation routing table 5O ₆ , forming, with the routing tables 5O ₅ of each personal network, a data transfer system allowing the exchange of data between the terminals made available to the federation. This federation routing table is based on the flow of invitation and acceptance messages. It groups together all source and destination personal network pairs, with, associated with each pair, the personal network to which the data is transferred to reach the destination.

It will be noted that the transfer of data is thus possible without recourse to subsequent message exchanges.

In the example illustrated, the federation routing table for the personal network 64 is expressed as:

68 60 60 641

66 60 60 641

60 68 68 642

The first column is the identifier of the source personal network, the second column is the identifier of the destination personal network, the third column is the identifier of the next personal network to which the data is to be transferred, and the fourth column is the local gateway allowing to reach the next personal network.

The federation routing table further includes, in a variant, costs associated with each subsequent personal network, in order to find the optimal personal network route when several choices are possible.

These costs are, for example, the sum of the costs of the different connections between terminals separating two personal networks. Thus, the cost to reach the personal network 60 from the personal network 68, through the personal network 64 would be the accumulation of the costs of the links between the gateway 681 of the personal network 68 and the gateway 6O1, through the gateways 64 ₂ and 64 ₁ of the network 64.

Thus, the data can be exchanged between personal networks using the federation routing table 5O ₆ and, in a personal network, its own routing table 5O ₅ . With reference to FIG. 6, the formation of a remote federation of personal networks will now be described.

In this example, four personal networks 70, 72, 74 and 76 are involved. These personal networks are interconnected by the Internet network 78 forming interconnection network. The interconnection gateways between the personal networks and the Internet are indicated in hatched lines. They are referenced 70-ι, 72 _1; 74 _1; 76-ι. Furthermore, each personal network is associated with its respective agent 7O ₂ , 72 ₂ , 74 ₂ , 76 ₂ in the Internet network 78. As indicated above, the agents are intended to allow the exchange of data between sectors of a network. same personal network.

In the invention, the various agents 7O ₂ , 72 ₂ , 74 ₂ , 76 ₂ are further configured to communicate with one another. This is for example achieved by establishing one or more listing servers, listing the addresses of the agents, each associated with the identifier of the corresponding personal network.

The owner of the personal network 70 decides to form a federation of personal networks, for example to launch an online video game, or to share vacation photos. The owner subsequently uses any of the terminals of his personal network 70.

The owner creates, during a step 300, a federation profile.

The federation profile includes an identifier of its creator, that is to say the identifier of the personal network of the creator, a textual description indicating the purpose of the federation, access control policies, resources or services made available or required for the federation, as well as a list of participants identified by the identifier of their personal network. For example, the list of participants is personal networks 74 and 76.

The owner of the personal network 70 then creates, during a step 302, a participation profile in which he notably defines the resources of his personal network 70 that he makes available to the federation. During a step 304, the two profiles, of federation and participation, are made available to the local gateways of the creator's personal network, in the example, it is the local gateway 701.

The federation profile is then distributed to the participants' personal networks.

To this end, the federation's profile is transmitted to the personnel officer 7O ₂ network creator 70 during a step 306, then the agent automatically transmits the federation profile to agents personal networks included as a participant in profile federation, thanks to their identifier.

Thus, each participating personal network receives the federation profile and presents it to its owner (step 308).

If the latter accepts the invitation, the owner of the personal network creates (step 310) a participation profile indicating the resources of his personal network that he makes available to the federation and transmits (step 312) a return request accompanied by the Participation profile for the creator's personal network, through his agent.

The resources indicated in the participation profile specify the IP address of the terminals sharing their resources.

Thus, the creator's personal network selects terminals among those made available by the different participants and modifies the federation profile to include the list of selected terminals, as well as their IP address.

This modified federation profile is distributed again to the participants' personal networks.

Preferably, the modified federation profile is stored in each of the personal network agents of the participants. Thus, each personal network has at its disposal, in its associated agent, the IP address of each of the terminals of the federation. The agents 70 ₂ ... 76 ₂ thus form a data transfer system enabling the exchange of data between the terminals of the federation using a conventional IP routing provided by the interconnection network. Network connectivity is established.

It will be noted that the invention is not limited to the embodiment described. In particular, the creator of a local federation may not specify a list of potential participants. In this case, the invitation message is broadcast step by step, for example until a predefined number of participants is reached, or until a limit in number of network hops is reached.

Claims

A routing method for federating a plurality of networks (60 ... 68; 70 ... 76), the plurality of networks comprising at least two interconnection gateways between two networks, each network comprising a topology own terminal represented by an associated network routing table (5O ₅ ) and network operable resources, the method comprising:

a step of generating a federation routing table (5O ₆ ) by said at least one network, a federation routing table (5O ₆ ) comprising a federation topology according to the clean topologies of the networks having adhered to the federation and choice of common resources operated by said networks, characterized in that the step of developing a federation routing table (5O ₆ ) takes into account messages exchanged during the accession step, without use of subsequent message exchanges.

2. Method according to claim 1, characterized in that it comprises the steps of:

- developing a resource pool invitation message including an inviting network identifier (60; 70),

sending the invitation message from the inviting network (60; 70) to each other guest network (62 ... 68; 72 ... 76) of the plurality,

- following receipt of the invitation message by a guest network (60 ... 68; 72 ... 76), elaborate an acceptance message for the pooling of resources including at least the address of a terminal of this guest network,

sending the acceptance message from at least one guest network (62 ... 68, 72 ... 76) to the inviting network (60; 70),

selecting at least one terminal of the inviting network and at least one terminal of the guest network contained in the acceptance message or messages, in order to pool all or part of the resources of these terminals, sending to the networks (60 ... 68, 70 ... 76) of the plurality, a pooling message including the list of selected terminals,

developing, on the basis of data conveyed by at least part of the invitation, acceptance and pooling messages, a data transfer system (5O ₅ , 5O ₆ , 70 ₂ ... 76 ₂ ) the exchange of data between the selected terminals, without recourse to subsequent message exchanges.

3. Method according to claim 2, characterized in that at least two networks (70 ... 76) of the plurality are distant from each other, separated by an interconnection network (78), and the process comprises the following steps:

configuring, in the interconnection network (78), a first agent (70 ₂ ... 76 ₂ ) associated with the first remote network and a second agent associated with the second remote network, so that each network and its associated agent can transfer data between them,

performing the exchange of messages between the two remote networks via the agents (70 ₂ ... 76 ₂ ),

developing the data transfer system by making available to each network the addresses of the selected terminals provided by the acceptance messages, in order to be able to use a routing provided by the interconnection network (78).

4. Method according to claim 3, characterized in that data are exchanged between at least two sectors of a network thanks to the associated agent, the sectors being distant from one another, separated by the network of interconnection.

5. Method according to claim 3 or 4, characterized in that it comprises the step of transferring, to each agent (70 ₂ ... 76 ₂ ), the list of selected terminals.

6. Method according to claim 2, characterized in that networks (60 ... 68), forming at least one subset of the plurality of re- and networks comprising the inviting network, are directly connected to each other, and in that the step of developing a data transfer system comprises the development, by each network of the set, of a routing table (5O ₅ ) among the networks of the set, from the conveyed data consisting of the routing in the subset of the invitation and acceptance messages.

7. Method according to claim 6, characterized in that the invitation message is sent to the networks of the set (62 ... 68) step by step starting from the inviting network (60).

8. Process according to claim 7, characterized in that:

the invitation message includes a list of networks of the set,

9. Method according to any one of claims 6 to 8, characterized in that the pooling message comprising the list of selected terminals is stored in each terminal of each network of the assembly, forming a gateway (6O ₁ , 62 - ,, 64- ,, 64 ₂ , 66 _1; 66 ₂ , 68-ι) with another network of the set.

10. Method according to any one of the preceding claims, characterized in that each terminal of each network of the plurality establishes a local routing table, in which each directly connected terminal is associated with the IP address of the terminal and with least one MAC address of a network interface to which this terminal is directly connected.

1 1. Process execution software defined according to any one of claims 1 to 10.

A network federation system comprising a plurality of networks (60 ... 68; 70 ... 76), each network comprising an own topology of terminal nodes represented by a network routing table. associated (5O ₅ ) and the plurality of networks comprising at least two interconnection gateways between two networks, characterized in that it comprises means for generating a federation routing table (5O ₆ ), for at least a network having joined the federation according to a predetermined approval protocol, said federation routing table (5O ₆ ) comprising a federative topology according to the proper topologies of the networks having joined the federation and the choices of common resources operated by said networks , and in that the means for generating a federation routing table (5O ₆ ) are configured to develop the routing table from messages exchanged during the execution of the approval protocol, without recourse to subsequent message exchanges.

Terminal intended to be selected in a federation of networks of a plurality and belonging to a network of the plurality comprising a routing engine (50) characterized in that it comprises a module (50-ι) of terminal discovery adjacent a module (5O ₂₎ discovery of adjacent networks, a module (5O ₃₎ federation management, a module (5O ₄₎ authentication, a routing table (5O ₅₎ for the network to which the terminal belongs considered, and a routing table (5O ₆ ) for the network federation of which said terminal belongs.